Prognostic Impact of the SARC-F Score in Gastrointestinal Advanced Cancers

Simple Summary There have been few reports with regard to the relevance between the SARC-F score and the prognosis in patients with gastrointestinal advanced cancers, and we aimed to elucidate these issues (n = 421, median age = 73 years). During the follow-up period, 145 patients (34.4%) died. The 1-year cumulative overall survival rate in patients with SARC-F ≥ 4 (recommended cutoff point, n = 103) and SARC-F < 4 (n = 318) was 33.9% and 61.6% (p < 0.0001). In the multivariate analysis for the overall survival, total lymphocyte count ≥ 1081/μL (p = 0.0014), the SARC-F score ≥ 4 (p = 0.0096), Glasgow prognostic score 1 (p = 0.0147) and 2 (p < 0.0001), ECOG-PS 2 (p < 0.0001), and 3 (p < 0.0001) and 4 (p < 0.0001) were independent predictors. In the receiver operating characteristic curve analysis on the prognostic value of the SARC-F score, the sensitivity/specificity was 0.59/0.70, and the best cutoff point of the SARC-F score was two. The SARC-F score appears to be useful in patients with gastrointestinal advanced malignancies. Abstract We sought to elucidate the prognostic impact of the SARC-F score among patients with gastrointestinal advanced malignancies (n = 421). A SARC-F score ≥ 4 was judged to have a strong suspicion for sarcopenia. In patients with ECOG-PS 4 (n = 43), 3 (n = 61), and 0–2 (n = 317), 42 (97.7%), 53 (86.9%) and 8 (2.5%) had the SARC-F score ≥ 4. During the follow-up period, 145 patients (34.4%) died. All deaths were cancer-related. The 1-year cumulative overall survival (OS) rate in patients with SARC-F ≥ 4 (n = 103) and SARC-F < 4 (n = 318) was 33.9% and 61.6% (p < 0.0001). In the multivariate analysis for the OS, total lymphocyte count ≥ 1081/μL (p = 0.0014), the SARC-F score ≥ 4 (p = 0.0096), Glasgow prognostic score (GPS) 1 (p = 0.0147, GPS 0 as a standard), GPS 2 (p < 0.0001, GPS 0 as a standard), ECOG-PS 2 (p < 0.0001, ECOG-PS 0 as a standard), ECOG-PS 3 (p < 0.0001, ECOG-PS 0 as a standard), and ECOG-PS 4 (p < 0.0001, ECOG-PS 0 as a standard) were independent predictors. In the receiver operating characteristic curve analysis on the prognostic value of the SARC-F score, the sensitivity/specificity was 0.59/0.70, and best cutoff point of the SARC-F score was two. In conclusion, the SARC-F score is useful in patients with gastrointestinal advanced malignancies.


Introduction
Sarcopenia is characterized by generalized muscle mass loss and muscle functional decline, leading to physical frailty, cachexia, and mortality [1,2]. Malnutrition, reticence, advanced malignancy burden, and persistent inflammatory condition often identified in patients with gastrointestinal disorders are typical clinical characteristics associated with sarcopenia [1][2][3][4][5][6][7][8][9]. Patients with advanced cancer are prone to muscle protein loss due to muscle proteolysis-inducing factors and inflammatory cytokines released by the cancer cells [10]. In addition, psychological anxiety and side effects of treatment for the underlying malignancies can lead to a loss of appetite, and numerous previous studies have reported that sarcopenia is an adverse predictor in patients with advanced cancer [11][12][13]. Cancer cachexia is a metabolic abnormality with a systemic inflammatory response, and it is associated with increased catabolism, such as insulin resistance and increased lipid degradation, as well as increased skeletal muscle degradation [10]. In Japan, anamorelin, a ghrelin (appetite stimulating peptide) receptor agonist, is currently the only drug approved for cancer cachexia [14].
SARC-F (strength (S), assistance walking (A), rising from a chair (R), climbing stairs (C), and falls (F)) is a self-reported questionnaire consisting of five queries with the aim of screening for sarcopenia [15,16]. Subjects are asked to answer to the five questions on a scale of 0 to 2, and the total score is calculated [15]. A SARC-F score ≥ 4 is judged to have a strong suspicion for sarcopenia [15]. The SARC-F score can be linked to physical functional reserve and activity of daily life [17,18]. The SARC-F is recommended for use as an initial screening method for sarcopenia in the current international guidelines [19,20], whereas its lower sensitivity for sarcopenia may be a problem [21][22][23]. Therefore, attempts have been made to combine the SARC-F score with age, body mass index (BMI), and calf-circumference to increase sensitivity for sarcopenia [24][25][26]. The use of SARC-calf for sarcopenia screening is also recommended in the current Asian guidelines [19].
To our knowledge, however, there have been few reports with regard to the relevance between the SARC-F score and prognosis in patients with gastrointestinal advanced cancers [27]. Such research questions may deserve to be clarified, which urged us to perform the present analysis.

Patients and Our Study
In the Second Department of Internal Medicine, Osaka Medical and Pharmaceutical University (OMPU) hospital, sarcopenia risk has been assessed by the SARC-F. As a rule, all hospitalized patients were asked to answer the SARC-F questionnaire upon admission. Between May 2020 and September 2021, 421 advanced malignancy patients with SARC-F scores can be found in our database. Stage III or more severe cancer was defined as advanced cancer. In each patient, the most appropriate therapy was chosen through discussion with gastroenterologists, surgeons, and oncologists based on the current guidelines in each cancer. The primary outcome measure was overall survival (OS). First, the OS rate was compared according to the SARC-F score. Next, univariate and multivariate analyses for the OS were performed. Next, survival analysis was performed according to the disease primary site. Finally, a receiver operating characteristic (ROC) curve analysis for the SARC-F based on the prognosis score was performed.

Glasgow Prognostic Score (GPS)
GPS is calculated by the two serum markers: serum albumin level (reference value = 3.5 g/dL) and C reactive protein (CRP) level (reference value = 1.0 mg/dL), and it is a frequently used prognostic system in patients with advanced malignancies [28,29]. GPS 0 was determined as patients with a serum albumin level > 3.5 g/dL and CRP < 1.0 mg/dL, GPS 2 those with serum albumin level < 3.5 g/dL and CRP >1.0 mg/dL. The remaining subjects were classified as GSP 1.

Statistics
The relationship between two variables was assessed by Pearson correlation coefficient (r). For the analysis of the significance of prognostic parameters, continuous parameters were dichotomized at the median value. Kaplan-Meier method was used for the cumulative Cancers 2022, 14, 10 3 of 10 OS rate and the log-rank test was used for testing. Parameters with a p value < 0.05 in the univariate analysis were entered into the multivariate analysis (Cox proportional hazard model). The follow-up period was defined as the period from the date of SARC-F score acquisition to the date of confirmed death or last confirmed survival. ROC analysis based on the prognosis was also done for the area under the ROC (AUC) of the SARC-F score, and the cutoff was adopted as the point where the sum of sensitivity and specificity is maximized. In the data presentation, median value (interquartile range (IQR)) was used. A p = 0.05 was set at the significant level by the JMP ver. 15 (SAS Institute Inc., Cary, NC, USA).

ROC Analysis on the Prognstic Impact of the SARC-F Score
In the ROC analysis on the prognostic impact of the SARC-F score, the AUC was 0.66 ( Figure 4A). The sensitivity/specificity was 0.59/0.70, and the optimal cutoff point of the SARC-F score was 2 ( Figure 4A). When the cutoff point of the SARC-F score was changed to 2, the 1-year cumulative OS rate in patients with the SARC-F score ≥ 2 (n = 132) and <2 (n = 289) was 31.6% and 65.0%, respectively (p < 0.0001, Figure 4B).

ROC Analysis on the Prognstic Impact of the SARC-F Score
In the ROC analysis on the prognostic impact of the SARC-F score, the AUC was 0.66 ( Figure 4A). The sensitivity/specificity was 0.59/0.70, and the optimal cutoff point of the SARC-F score was 2 ( Figure 4A). When the cutoff point of the SARC-F score was changed to 2, the 1-year cumulative OS rate in patients with the SARC-F score ≥ 2 (n = 132) and <2 (n = 289) was 31.6% and 65.0%, respectively (p < 0.0001, Figure 4B).

Discussion
In Japan, due to the spread of the COVID-19 infection, there is a concern that the number of patients receiving cancer screening will decrease and the number of cancer patients who are found in an advanced stage will increase. Although the number of new COVID-19 infections has been decreasing due to the widespread use of COVID-19 vac-

Discussion
In Japan, due to the spread of the COVID-19 infection, there is a concern that the number of patients receiving cancer screening will decrease and the number of cancer patients who are found in an advanced stage will increase. Although the number of new COVID-19 infections has been decreasing due to the widespread use of COVID-19 vaccines, there is a concern that people, especially the elderly, may be reluctant to receive medical check-ups in a hospital. Thus, an easier screening method without an invasive procedure for the medical check-up may be ideal. As described above, the recommendation in the current international guidelines involves the use of the SARC-F score as an initial screening tool for sarcopenia [12,19,20]. However, to our knowledge, few reports can be found regarding the SARC-F score and the prognosis in patients with advanced gastrointestinal malignancies. If the prognostic impact of SARC-F is clarified, it will be easy to estimate the prognosis of patients with advanced cancer. This is especially significant in facilities with difficulties for assessing muscle mass. We thus conducted the current analysis. We believe that the current results are worth reporting.
In our results, a SARC-F score ≥ 4 was an independent factor for the OS, and in all subgroup analyses except for HCC, patients with a SARC-F score ≥ 4 had significantly worse survival than those with a SARC-F score < 4. These results denoted that the SARC-F is a helpful prognostic scoring system in patients with gastrointestinal advanced malignancies. The reason for the non-significant difference of the OS between the patients with the SARC-F score ≥ 4 and <4 in HCC patients may be attributed to the distribution of the GPS (GPS 0/GPS 1/GPS 2: 5 (41.7%)/5 (41.7%)/2 (16.7%) in the SARC-F score ≥ 4, and 18 (41.9%)/13 (30.2%)/12 (27.9%) in the SARC-F score < 4). A previous meta-analysis demonstrated the prognostic impact of the GPS in HCC patients [30].
In our data, there was a strong correlation between the SARC-F score and ECOG-PS (r = 0.89). In patients with PS, 4 (n = 43), 42 (97.7%) had a SARC-F score ≥ 4. In patients with PS 3 (n = 61), 53 (86.9%) had the SARC-F score ≥ 4. While in patients with PS 0-2 (n = 317), only 8 (2.5%) had a SARC-F score ≥ 4. The SARC-F is a questionnaire involving 5 queries as mentioned in the introduction section. The ECOG-PS indicates the degree of limitation in the patient's daily life. The SARC-F can be an alternative for ECOG-PS. While in our multivariate analysis, total lymphocyte count was an independent predictor as well as EOCG-PS, GPS, and the SARC-F score. The total lymphocyte count reflects immunological function [31]. Exercise can ameliorate sarcopenia through the improvement of immunological function, and thus, exercise can be recommended in patients with advanced cancer [32][33][34]. GPS is a well-established prognostic marker among patients with advanced malignancies [28,29,35]. The ORs for the OS of GPS 1 and GPS 2 in this study were 1.953 (p = 0.0147, GPS 0 as a reference) and 3.460 (p < 0.0001, GPS 0 as a reference), which were consistent with previous reports [28,29,35]. Although BMI was not significant for the OS in our results, body weight loss was an important factor associated with cancer cachexia [10]. In our cohort, 86 patients (20.4%) had BMI < 18.5 kg/m 2 (underweight). In such patients, appropriate nutritional intervention may be required to improve prognosis.
In the ROC analysis based on the prognosis for the SARC-F score, AUC was 0.66, the sensitivity/specificity was 0.59/0.70, and the optimal cutoff point of the SARC-F score was 2. When the cutoff point of the SARC-F score was set at 4 (recommended cutoff point), the sensitivity/specificity was 0.38/0.83, indicating a significant decrease in the sensitivity. As described above, a relatively lower sensitivity of the SARC-F score for sarcopenia may be a problem, which is identical to our data [21][22][23]. Hanai et al. reported in their large cohort with chronic liver disease that a SARC-F score = 1 demonstrated a stronger discrimination capability for selecting sarcopenia than a SARC-F score = 4, with an AUC of 0.68, and with the sensitivity/specificity of 0.65/0.68 [36]. In view of these, the cutoff point of the SARC-F score should be carefully reviewed.
We should mention several limitations to the current analysis. First, the study involved a single-center retrospective observational study design. Second, the exact number of patients with a definite diagnosis of sarcopenia was unclear in our data. Third, our study Cancers 2022, 14, 10 8 of 10 cohort was highly heterogeneous, including a broad spectrum of gastrointestinal advanced malignancies, also creating bias. Fourth, the SARC-F score can vary depending on the patient status, such as malnutrition and dehydration. The patient's prognosis should therefore be evaluated comprehensively, taking into account factors other than the SARC-F score. Finally, the relatively short follow-up period (median, 0.4 years) in our study should also be mentioned, but the number of deaths was 145 in this study, which can be appropriate for survival analysis.

Conclusions
The SARC-F score in patients with gastrointestinal advanced malignancies efficiently reflects survival. Appropriate therapeutic interventions should be considered in patients with gastrointestinal advanced malignancies with a higher SARC-F score.